This chapter covers Amazon S3 storage classes, from S3 Standard to Glacier Deep Archive. Understanding storage classes is critical for the CLF-C02 exam because questions on cost optimization, data lifecycle management, and retrieval trade-offs appear frequently, contributing approximately 8% to the 'Cloud Technology Services' domain. You will learn the specific characteristics, use cases, and pricing models of each class, as well as how to transition data between them using lifecycle policies.
Jump to a section
Imagine you run a business that stores documents for clients. You have a large self-storage warehouse with different zones based on how quickly clients need access. The main floor (S3 Standard) is where you keep documents that clients might request at any moment—high rent, but instant retrieval. For documents that are rarely needed but must be available within minutes, you use a back room (S3 Standard-IA); rent is lower, but you pay a fee each time you retrieve a box. For archives you almost never touch, you have a basement (S3 Glacier) where boxes are stored on shelves; retrieval takes hours because a staff member must locate and bring the box up, and there's a fee per retrieval. For deep archives that are legally required to be kept for years but never accessed, you have a sealed vault (S3 Glacier Deep Archive) where retrieval takes 12 hours or more and retrieval fees are high. Amazon S3 storage classes work exactly like this: you pay less for storage the less you need access, but you pay retrieval costs and wait longer when you do need the data. The mechanism is that AWS uses different underlying storage media, redundancy schemes, and retrieval workflows to optimize cost vs. latency.
What Are S3 Storage Classes and Why Do They Exist?
Amazon Simple Storage Service (S3) is a highly durable object storage service designed to store and retrieve any amount of data from anywhere. However, not all data is accessed with the same frequency or urgency. Storing all data in the highest-performance tier would be unnecessarily expensive for infrequently accessed or archival data. S3 storage classes address this by offering different tiers optimized for various access patterns, balancing cost, availability, and retrieval time.
The core problem S3 storage classes solve is the trade-off between storage cost and retrieval cost. Data that is accessed frequently needs to be stored on high-performance media with low latency and high throughput, which costs more per gigabyte. Data that is rarely accessed can be stored on lower-cost media (like tape or slower disks) but incurs higher retrieval costs and longer wait times. S3 storage classes automate this choice, allowing you to define lifecycle policies that move data between classes automatically based on age or access patterns.
How S3 Storage Classes Work
Each S3 object (file) is stored in a bucket and assigned a storage class at the object level. The storage class determines: - Durability: All S3 classes offer 99.999999999% (11 nines) durability, meaning data is replicated across multiple devices and facilities. - Availability: The percentage of time the object is accessible for requests. Standard offers 99.99% availability, while Standard-IA offers 99.9%. - Minimum storage duration: Some classes require a minimum storage period (e.g., 30 days for Standard-IA, 90 days for Glacier) to avoid early deletion fees. - Retrieval time: From milliseconds (Standard) to minutes (Glacier Flexible Retrieval) to hours (Glacier Deep Archive). - Cost per GB stored: Decreases as you move to colder classes. - Cost per request: Increases for colder classes, especially for data retrieval.
AWS uses different underlying infrastructure for each class. For example, S3 Standard uses SSDs or fast HDDs with multiple replicas in different Availability Zones (AZs) within a region. S3 Glacier uses tape libraries or slower HDDs with a retrieval process that requires moving data to a staging area. S3 One Zone-IA stores data in a single AZ, reducing cost but also reducing availability if that AZ fails.
Key Storage Classes and Their Characteristics
S3 Standard: Designed for frequently accessed data. Retrieval time: milliseconds. Availability: 99.99%. No minimum storage duration. No retrieval fees. Use cases: active websites, content distribution, big data analytics.
S3 Intelligent-Tiering: Automatically moves data between access tiers based on changing access patterns. There are three access tiers: frequent, infrequent, and archive instant. You pay a monthly monitoring fee per object but no retrieval fees. Best for data with unknown or unpredictable access patterns.
S3 Standard-IA (Infrequent Access): For data accessed less frequently but requiring rapid access when needed. Lower storage cost than Standard, but you pay a retrieval fee per GB retrieved. Minimum storage duration: 30 days. Use cases: backups, disaster recovery files.
S3 One Zone-IA: Same as Standard-IA but stored in a single AZ. Lower cost, but data is lost if the AZ fails. Availability: 99.5%. Use cases: non-critical, reproducible data.
S3 Glacier Instant Retrieval: For archive data that needs immediate access (milliseconds). Similar cost to Standard-IA but with a 90-day minimum storage duration. Use cases: medical images, news media assets.
S3 Glacier Flexible Retrieval (formerly Glacier): For archival data that can tolerate retrieval times of minutes to hours. Three retrieval options: Expedited (1-5 minutes, higher cost), Standard (3-5 hours), Bulk (5-12 hours). Minimum storage duration: 90 days. Use cases: long-term backups, digital preservation.
S3 Glacier Deep Archive: Lowest-cost storage for data that is accessed at most once or twice per year. Retrieval times: Standard (12 hours), Bulk (48 hours). Minimum storage duration: 180 days. Use cases: regulatory archives, tape replacement.
S3 Outposts: Storage class for on-premises S3 on AWS Outposts. Not typically covered on CLF-C02.
Pricing Model
S3 pricing is composed of: - Storage price: Per GB per month, decreasing from Standard to Deep Archive. - Request and data retrieval pricing: Per request (PUT, GET, etc.) and per GB retrieved (for IA and Glacier classes). - Data transfer pricing: Data transfer IN to S3 is free; data transfer OUT to the internet is charged. - Lifecycle transition pricing: When you transition an object from one class to another, you pay a per-object fee (e.g., $0.01 per 1,000 transitions). - Early deletion fees: If you delete an object before the minimum storage duration, you pay the remainder of the storage cost as if you had kept it.
Example: Storing 1 TB of data in S3 Standard costs about $23/month. In S3 Glacier Deep Archive, it costs about $1/month. However, retrieving 1 TB from Deep Archive costs about $2.50 plus request fees, while Standard retrieval is free.
Comparison to On-Premises
In an on-premises environment, you would typically buy a single tier of storage (e.g., all SSDs or all HDDs) and accept either high cost or slow performance. With S3 storage classes, you can mix tiers within the same bucket, automatically moving data based on policies. This is more cost-effective and scalable than managing multiple storage systems manually.
When to Use Each Class
Standard: Primary data, active datasets.
Intelligent-Tiering: When you don't know access patterns.
Standard-IA: Backups that are rarely restored.
One Zone-IA: Temporary data, replicated elsewhere.
Glacier Instant Retrieval: Archives that need instant access.
Glacier Flexible Retrieval: Long-term backups, compliance.
Glacier Deep Archive: Legal holds, data that may never be accessed.
Lifecycle Policies
Lifecycle policies automatically transition objects between storage classes or delete them based on age. For example, you can move objects to Standard-IA after 30 days, to Glacier after 90 days, and delete after 365 days. Policies apply to a bucket or a prefix. The exam often tests that you can set a lifecycle policy to move data to Glacier after a certain number of days.
Create an S3 Bucket
First, create an S3 bucket via the AWS Management Console, CLI, or SDK. Choose a globally unique name and select a region. You can set default encryption and block public access. For storage classes, you don't set a default class at bucket creation; you assign it per object. However, you can later create a lifecycle rule that applies to all objects. The bucket itself has no storage class.
Upload an Object with a Storage Class
When uploading an object, you explicitly choose the storage class. In the console, you see a dropdown labeled 'Storage class' with options like Standard, Intelligent-Tiering, Standard-IA, etc. If you don't choose, the default is S3 Standard. Via CLI, you use the `--storage-class` parameter, e.g., `aws s3 cp myfile.txt s3://mybucket/ --storage-class STANDARD_IA`. Behind the scenes, AWS stores the object in the appropriate infrastructure and metadata indicates the class.
Set a Lifecycle Policy for Transitions
In the bucket's Management tab, create a lifecycle rule. Define a filter (prefix or tags) to select objects. Add a transition action: for example, 'Transition to Standard-IA after 30 days'. You can add multiple transitions, e.g., after 90 days move to Glacier. AWS applies the rule asynchronously; objects are checked daily. The transition incurs a per-object fee. Important: You cannot transition from Glacier to Standard directly; you must restore first.
Retrieve an Object from Glacier
To access an object stored in Glacier or Deep Archive, you must first initiate a restore request. This moves the object to a temporary copy in Standard or Standard-IA for a specified number of days. During restoration, you pay retrieval fees based on the retrieval tier (Expedited, Standard, Bulk). Once restored, you can download the object via GET. After the restore period, the temporary copy is deleted. The original object remains in Glacier.
Monitor Storage Costs with AWS Cost Explorer
Use AWS Cost Explorer to analyze storage costs by storage class. Filter by service 'Amazon S3' and group by 'Storage class'. This helps identify if you are overusing Standard for infrequent data. You can also set up billing alerts. The exam may ask about cost optimization strategies like using lifecycle policies to move data to lower-cost tiers.
Scenario 1: Media Company Storing Video Archives
A media company produces thousands of video files daily. Initially, videos are stored in S3 Standard for editing and quick access. After 30 days, a lifecycle policy moves them to S3 Standard-IA because they are rarely edited but may be referenced. After 90 days, they move to S3 Glacier Flexible Retrieval for long-term archiving. If a video needs to be reused, the team initiates a restore (Standard retrieval, 3-5 hours). This saves 80% on storage costs compared to keeping everything in Standard. Misconfiguration: If the team accidentally sets the lifecycle to delete instead of transition, they lose data. Also, if they need immediate access to a Glacier object, the 3-5 hour wait causes delays.
Scenario 2: Healthcare Compliance with Deep Archive
A hospital must retain patient records for 10 years per regulations. They store records in S3 Standard for the first year (active use), then transition to S3 Glacier Deep Archive. Retrieval is extremely rare (e.g., legal discovery). The low storage cost of Deep Archive ($1/TB/month) makes it affordable. However, if a retrieval is needed, it takes 12 hours and costs retrieval fees. The hospital must budget for these rare events. A common mistake is forgetting the 180-day minimum storage duration; if a record is deleted before 180 days, an early deletion fee applies.
Scenario 3: E-commerce Website with Intelligent-Tiering
An e-commerce site stores product images and logs. Access patterns are unpredictable: some images become popular, others never viewed. Using S3 Intelligent-Tiering automatically moves rarely accessed images to lower-cost tiers without manual intervention. The monthly monitoring fee ($0.0025 per 1,000 objects) is negligible. However, if the site has a sudden spike in traffic, Intelligent-Tiering may move objects back to frequent access tier, incurring transition fees. The team must monitor for unexpected costs.
What CLF-C02 Tests on S3 Storage Classes
This objective falls under 'Cloud Technology Services' (Domain 3) and specifically objective 3.2: 'Compare storage services and their use cases'. The exam expects you to:
Identify the appropriate storage class given a use case (e.g., 'frequently accessed data' -> S3 Standard).
Understand the trade-offs between cost, retrieval time, and durability.
Know the minimum storage durations: 30 days for Standard-IA and One Zone-IA, 90 days for Glacier Instant Retrieval and Glacier Flexible Retrieval, 180 days for Glacier Deep Archive.
Recognize that all S3 classes have 11 nines durability, but availability varies (99.99% Standard vs 99.9% Standard-IA).
Understand that lifecycle policies can transition objects between classes automatically.
Common Wrong Answers and Why Candidates Choose Them
'S3 Standard is best for backups because it's the cheapest.' Wrong because Standard is the most expensive. Candidates confuse Standard with Standard-IA or Glacier.
'Glacier Deep Archive has the fastest retrieval time.' Wrong because it has the slowest (12+ hours). Candidates think 'Deep Archive' implies deeper storage but faster? No.
'S3 One Zone-IA has 99.99% availability.' Wrong; it has 99.5% because it's in one AZ. Candidates assume all S3 classes have the same availability.
'You can directly read an object in Glacier without restoring.' Wrong; you must restore first. Candidates think Glacier is like a slow disk.
Specific Terms That Appear on the Exam
'Durability' vs 'Availability': Durability is about data loss; availability is about uptime.
'Lifecycle policy' vs 'Replication rule': Lifecycle transitions storage classes; replication copies objects.
'Retrieval fee' vs 'Storage fee': IA and Glacier have retrieval fees; Standard does not.
'Minimum storage duration': Early deletion fees apply if deleted before the period.
Tricky Distinctions
S3 Standard-IA vs S3 One Zone-IA: Both have 30-day minimum, but One Zone-IA is cheaper and less durable (single AZ).
S3 Glacier Instant Retrieval vs S3 Standard-IA: Both have millisecond retrieval, but Glacier Instant has a 90-day minimum and higher retrieval cost.
S3 Glacier Flexible Retrieval vs S3 Glacier Deep Archive: Flexible has 1-5 minute expedited retrieval; Deep Archive has 12-hour minimum.
Decision Rule for Multi-Choice Questions
Step 1: Identify the access frequency (frequent, infrequent, archive). Step 2: Identify retrieval time requirement (milliseconds, minutes, hours). Step 3: Identify cost sensitivity. Then match to the class. If 'unknown pattern', choose Intelligent-Tiering. If 'single AZ acceptable', choose One Zone-IA.
S3 storage classes are Standard, Intelligent-Tiering, Standard-IA, One Zone-IA, Glacier Instant Retrieval, Glacier Flexible Retrieval, and Glacier Deep Archive.
All S3 classes offer 99.999999999% durability (11 nines).
S3 Standard-IA and One Zone-IA have a 30-day minimum storage duration; Glacier classes have 90 or 180 days.
Lifecycle policies can automate transitions between storage classes based on object age.
You cannot directly read objects in Glacier or Deep Archive; you must restore them first.
S3 One Zone-IA stores data in a single Availability Zone, offering 99.5% availability.
S3 Intelligent-Tiering is best for unpredictable access patterns, but incurs a monitoring fee.
Retrieval fees apply for Standard-IA, One Zone-IA, and all Glacier classes.
Early deletion fees apply if you delete objects before the minimum storage duration.
S3 Glacier Deep Archive is the cheapest storage class, designed for data accessed once or twice per year.
These come up on the exam all the time. Here's how to tell them apart.
S3 Standard
Millisecond retrieval time
No retrieval fees
No minimum storage duration
99.99% availability
Higher storage cost ($0.023/GB/month)
S3 Glacier Deep Archive
12-hour retrieval time (Standard)
Retrieval fees apply ($0.02/GB)
180-day minimum storage duration
99.9% availability
Lowest storage cost ($0.001/GB/month)
Mistake
All S3 storage classes offer the same availability.
Correct
S3 Standard offers 99.99% availability, Standard-IA and Glacier classes offer 99.9% (or less for One Zone-IA at 99.5%). Durability is 11 nines for all, but availability differs.
Mistake
You can change the storage class of an object without any cost.
Correct
Transitioning objects between storage classes incurs a per-object fee (e.g., $0.01 per 1,000 transitions) and may incur retrieval fees if moving from a colder class.
Mistake
S3 Glacier Deep Archive is the same as S3 Glacier but slower.
Correct
Deep Archive has a different pricing model, minimum storage duration (180 days vs 90 days), and retrieval options (12 hours vs 1-5 minutes expedited). They are separate classes.
Mistake
S3 Intelligent-Tiering is free to use.
Correct
Intelligent-Tiering charges a monthly monitoring fee per object ($0.0025 per 1,000 objects) plus storage costs for each tier. It is not free.
Mistake
You can directly download an object from S3 Glacier without restoring.
Correct
Objects in Glacier or Deep Archive are not directly accessible via GET. You must initiate a restore request, which creates a temporary copy in Standard or Standard-IA for a specified duration.
S3 Standard is for frequently accessed data with millisecond retrieval and no retrieval fees, but higher storage cost. S3 Standard-IA is for infrequently accessed data with lower storage cost but charges a retrieval fee per GB. Both have 11 nines durability, but Standard has 99.99% availability while Standard-IA has 99.9%. Standard-IA also has a 30-day minimum storage duration; early deletion incurs a fee.
Yes, you can change the storage class of an existing object by copying it to the same bucket with a new storage class, or by using a lifecycle policy. However, transitioning between classes incurs a per-object fee. You cannot directly change from Glacier to Standard without restoring first.
S3 Glacier Deep Archive is the cheapest, at about $0.001 per GB per month (varies by region). However, retrieval times are long (12 hours for Standard) and retrieval fees are high. It is suitable for data that is rarely accessed and can tolerate long wait times.
You must initiate a restore request via the console, CLI, or SDK. Choose a retrieval tier: Expedited (1-5 minutes), Standard (3-5 hours), or Bulk (5-12 hours). The object is temporarily stored in Standard or Standard-IA for a specified number of days. During that period, you can download it. After the period, the temporary copy is deleted.
You are charged an early deletion fee equal to the remaining storage cost for the minimum duration. For example, if you delete a Standard-IA object after 10 days, you pay the storage cost for the remaining 20 days. This applies to Standard-IA, One Zone-IA, and all Glacier classes.
Intelligent-Tiering is cost-effective for data with unpredictable access patterns. However, it charges a monthly monitoring fee per object ($0.0025 per 1,000 objects). For data that is always active or always inactive, a fixed storage class may be cheaper. It also incurs transition fees when moving between tiers.
Yes, you can create a lifecycle rule that transitions objects to Glacier or Deep Archive after 0 days. However, you must consider the minimum storage durations and costs. Immediate transition may be useful for data that is known to be archival from the start.
You've just covered S3 Storage Classes — Standard to Glacier — now see how well it sticks with free CLF-C02 practice questions. Full explanations included, no account needed.
Done with this chapter?